AMD Radeon R5 M420 vs NVIDIA GeForce GT 640 Comparison

Head-to-Head Benchmarks

The only directly comparable benchmark between these two GPUs is Geekbench OpenCL, and the result is close. The AMD Radeon R5 M420 scores 3956, while the NVIDIA GeForce GT 640 scores 3736, giving AMD a 5.9% advantage. This is a narrow margin, and it places both cards in a similar performance tier.

Looking at the broader database context, the AMD Radeon R5 M420 sits at the 23rd percentile of all GPUs. Its nearest rivals include the NVIDIA GeForce 830M at 3957 (a 0% delta), the NVIDIA GeForce GT 745M at 3953 (0.1% ahead of AMD), the NVIDIA Quadro K2000 at 3964 (0.2% ahead), and the AMD Radeon HD 6850 X2 at 3977 (0.5% ahead). The AMD card is essentially locked in a dead heat with these four products, with deltas of half a percent or less in either direction. This indicates that the R5 M420 is positioned at the edge of a cluster of roughly equivalent mobile and older desktop parts.

The NVIDIA GeForce GT 640, by contrast, holds the 20th percentile of all GPUs and has an average benchmark score of 3210 across all tests. Its OpenCL result of 3736 is actually higher than its average, which is pulled down by additional benchmark runs. The nearest rivals for the GT 640 include the Intel Arc Pro B60 at 3182 (0.9% ahead of NVIDIA), the NVIDIA Quadro P1000 at 3163 (1.5% ahead), the NVIDIA GeForce 920M at 3287 (2.3% behind), and the NVIDIA GeForce GT 730M at 3316 (3.2% behind). The GT 640 therefore sits in a wider spread of competitors, with deltas ranging from roughly 1.5% to 3.2%.

When comparing the two cards directly, the AMD R5 M420 wins the only shared benchmark by 5.9%. That is a meaningful but not overwhelming gap in raw compute performance. In a head-to-head count, AMD takes 1 win and NVIDIA takes 0 wins. The data does not show any test where the GT 640 beats the R5 M420 in a directly comparable workload.

The difference in average benchmark scores is more pronounced. The R5 M420 averages 3956 across its benchmark suite, while the GT 640 averages 3210. That represents a roughly 23% gap in favor of AMD when considering the GT 640's full set of recorded tests. However, this comparison is complicated by the fact that the GT 640 has additional benchmark entries in Metal and Vulkan, which are not available on the AMD side. The GT 640 scores 2086 in Geekbench Metal and 3809 in Geekbench Vulkan, while the AMD R5 M420 has no recorded Metal or Vulkan results.

Where Each One Wins

The AMD Radeon R5 M420 wins in Geekbench OpenCL, the only direct test in the database for this pairing. This makes AMD the preferred choice for compute-oriented workloads that rely on OpenCL acceleration. The 5.9% lead suggests that the R5 M420 can handle general-purpose compute tasks with slightly more throughput than the GT 640.

The GT 640 wins in terms of available benchmark coverage, with three recorded tests spanning OpenCL, Metal, and Vulkan. Its OpenCL score of 3736 is lower than AMD's 3956, but the GT 640 also demonstrates capability in Metal with 2086 and in Vulkan with 3809. These are not directly comparable to AMD, since the R5 M420 has no Metal or Vulkan entries, but they indicate a broader software compatibility profile in the database.

For users prioritizing raw pixel throughput, the GT 640 has a higher pixel rate at 7.216 GPixel/s compared to the R5 M420's 6.800 GPixel/s. The texture rate also favors NVIDIA at 28.86 GTexel/s versus 17.00 GTexel/s. These theoretical figures suggest the GT 640 may handle fill-rate-bound tasks more effectively, even though the OpenCL score favors AMD.

In terms of floating-point performance, the GT 640 delivers 692.7 GFLOPS of FP32 compute, while the R5 M420 delivers 544.0 GFLOPS. Despite this 27% theoretical advantage for NVIDIA, the AMD card still posts the higher OpenCL score, which indicates that real-world compute performance does not always align with peak specification numbers.

The GT 640 also offers a wider memory bus at 128 bit versus 64 bit, and higher memory bandwidth at 28.51 GB/s versus 16.00 GB/s. The GT 640 has 2 GB of DDR3 memory, while the R5 M420 has 4 GB of DDR3. The larger capacity on the AMD side may be beneficial for workloads with larger working sets, but the NVIDIA card has more than 78% higher bandwidth.

FAQ

Q: Which GPU scores higher in Geekbench OpenCL?

A: The AMD Radeon R5 M420 scores 3956, while the NVIDIA GeForce GT 640 scores 3736, giving AMD a 5.9% advantage.

Q: How does the AMD Radeon R5 M420 compare to its nearest rivals?

A: The R5 M420 is within 0.5% of the NVIDIA GeForce 830M (3957), the NVIDIA GeForce GT 745M (3953), the NVIDIA Quadro K2000 (3964), and the AMD Radeon HD 6850 X2 (3977). The deltas range from 0% to 0.5% in either direction.

Q: What is the average benchmark score for each GPU?

A: The AMD Radeon R5 M420 has an average benchmark score of 3956, while the NVIDIA GeForce GT 640 has an average of 3210 across its recorded tests.

Q: Does the NVIDIA GeForce GT 640 have any benchmark advantages?

A: The GT 640 has recorded scores in Metal (2086) and Vulkan (3809), while the R5 M420 has no Metal or Vulkan entries. The GT 640 also holds the 20th percentile of all GPUs, while the R5 M420 holds the 23rd percentile.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce GT 640 has 384 shading units, while the AMD Radeon R5 M420 has 320 shading units.

Q: What is the memory configuration difference?

A: The AMD Radeon R5 M420 has 4 GB of DDR3 on a 64 bit bus with 16.00 GB/s bandwidth. The NVIDIA GeForce GT 640 has 2 GB of DDR3 on a 128 bit bus with 28.51 GB/s bandwidth.

Specification Differences

The two GPUs differ across nearly every specification category. The AMD Radeon R5 M420 uses the Jet chip with GCN 1.0 architecture and belongs to the Gem System (R5 M400) generation. The NVIDIA GeForce GT 640 uses the GK107 chip with Kepler architecture and belongs to the GeForce 600 generation. Both are built on a 28 nm process at TSMC, but the transistor counts differ substantially. AMD packs 690 million transistors on a 56 mm² die, giving a transistor density of 12.3M per mm². NVIDIA packs 1,270 million transistors on a 118 mm² die, giving a density of 10.8M per mm².

Clock behavior differs as well. The AMD side lists a base clock of 780 MHz and a boost clock of 850 MHz, with memory clocked at 1000 MHz (2 Gbps effective). The NVIDIA side lists no base or boost clock but has a memory clock of 891 MHz (1782 Mbps effective).

Memory capacity and bandwidth diverge sharply. AMD offers 4 GB of DDR3 on a 64 bit bus with 16.00 GB/s bandwidth. NVIDIA offers 2 GB of DDR3 on a 128 bit bus with 28.51 GB/s bandwidth. The NVIDIA card has more than twice the bus width and nearly double the bandwidth, while AMD has double the capacity.

Compute unit counts differ. The R5 M420 has 320 shading units, 20 TMUs, and 8 ROPs. The GT 640 has 384 shading units, 32 TMUs, and 16 ROPs. The pixel rate is 6.800 GPixel/s for AMD and 7.216 GPixel/s for NVIDIA. The texture rate is 17.00 GTexel/s for AMD and 28.86 GTexel/s for NVIDIA. FP32 performance is 544.0 GFLOPS for AMD and 692.7 GFLOPS for NVIDIA.

The power profile is different. The AMD R5 M420 is listed as an IGP with no TDP, power connectors, or suggested PSU. The NVIDIA GT 640 has a TDP of 65 W, a single-slot form factor, no power connectors, and a suggested PSU of 250 W. The GT 640 is 145 mm (5.7 inches) long.

The bus interface differs. AMD uses PCIe 3.0 x8, while NVIDIA uses PCIe 3.0 x16. Display outputs are portable-device dependent on the AMD side, while the GT 640 lists 1x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The NVIDIA card also has a launch MSRP of 99 USD.

Architecture Differences

The AMD Radeon R5 M420 is built on GCN 1.0, AMD's first-generation Graphics Core Next architecture, implemented on the Jet chip. This architecture was designed for efficient compute and graphics in a compact die. The R5 M420 belongs to the Gem System (R5 M400) generation, with a predecessor in the Solar System family and a successor in Polaris Mobile. It supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170.

The NVIDIA GeForce GT 640 is built on Kepler, NVIDIA's architecture that succeeded Fermi. It uses the GK107 chip and belongs to the GeForce 600 generation, with a predecessor in GeForce 500 and a successor in GeForce 700. It supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The Vulkan version is slightly newer on the NVIDIA side, while the DirectX feature level is slightly higher on the AMD side.

The manufacturing approach differs in density. AMD achieves 12.3 million transistors per square millimeter on its 56 mm² die, a higher density than NVIDIA's 10.8 million per square millimeter on a 118 mm² die. The NVIDIA die is more than twice the area and carries nearly twice the transistor count, which reflects a fundamentally different scaling strategy. Kepler was designed with a larger shader array and wider memory interface, while GCN 1.0 on Jet prioritized a smaller footprint suitable for integrated use in portable devices.

The absence of base and boost clocks on the NVIDIA side means the GT 640 runs at a fixed operating point, while the AMD side has a defined boost behavior from 780 MHz to 850 MHz. The GT 640 compensates with more shading units and higher theoretical throughput. The R5 M420's IGP classification indicates it is designed to be integrated into a portable system, while the GT 640 is a discrete single-slot card with its own display outputs and power requirements.

The architectural differences translate into distinct performance characteristics. The AMD card posts a higher OpenCL score despite lower peak FP32 and lower fill rates. The NVIDIA card offers a wider memory bus and more than double the texture rate. Each architecture makes different trade-offs, and the benchmark data reflects that the AMD design extracts more OpenCL performance from a smaller, lower-power configuration.